He spent a decade supporting CUDA despite much criticism and share price losses only to ultimately be proven right. If that isn’t a long term bet, what is?